PRODUCTIVITY IMPROVEMENT AND COST OPTIMIZATION OF SMALL AND
MEDIUM SCALE ENTERPRISES
by
UNMESH VISHWAS TAMHANKAR
Presented to the Faculty of the Graduate School of
The University of Texas at Arlington in Partial Fulfillment
of the Requirements
for the Degree of
MASTER OF SCIENCE IN INDUSTRIAL ENGINEERING
THE UNIVERSITY OF TEXAS AT ARLINGTON
MAY 2017
iii
Acknowledgements
I would like to thank Dr. Aera LeBoulluec for her constant support and motivation
during the last 12 months. Originally, I did not plan on pursuing the Thesis option for my
graduate coursework. It was in a meeting with Dr. LeBoulluec in the summer of 2016,
while discussing my doctoral prospective, that she encouraged me to pursue the thesis
option and generously agreed to be my faculty advisor for the thesis. Since then, she has
been instrumental in improving my understanding of the field.
She has always given me the freedom to pursue my ideas and had mentored me
so that I could be able to develop them in a structured manner. She has provided me with
all the necessary resources to facilitate a smooth progress. She has taught me, through
her own actions, professionalism and punctuality, which I will imbibe into my persona.
With her support, guidance and mentorship, I am a better engineer now than I was a year
ago.
Finally, I would like to thank my parents for being with me despite being half the
world away. They have believed in me regardless of the circumstance and they have
given me all the resources at their disposal so that I can have the future they always
wanted for me.
APRIL 19, 2017
iv
ABSTRACT
PRODUCTIVITY IMROVEMENT AND COST OPTIMIZATION OF SMALL AND MEDIUM
SCALE ENTERPRISES
Unmesh Vishwas Tamhankar, MS
The University of Texas at Arlington, 2017
Supervising Professor: Aera LeBoulluec
Productivity improvement and cost optimization have been a topic of discussion
for at least a century. The publication of Fredrick Taylor’s book ‘The Principles of
Scientific Management’ in 1911 can be considered as the beginning of the Efficiency
movement. Surprisingly, the awareness about the principles of Industrial Engineering is
relatively low in many family-owned Small and Medium Scale Enterprises (SMEs).
According to a World Bank report, formal SMEs make up 45% of total employment and
contribute 33% of GDP in emerging markets. In the USA, SMEs make up for more than
95% of all firms and employ 50% of private sector employees. Therefore, it is self-evident
that improvement in the productivity of this sector will have a sizable impact on the
economy.
Productivity increase and cost optimization are critical tasks for SMEs. In a
competitive environment, increasing productivity without installing additional capacity
enables the SMEs to avoid heavy investment and maximize their profitability. In
comparison with large-scale industries, the SMEs have less cash to spend, and skilled
professionals who can implement productivity improvement strategies such as Lean, Six
Sigma, Total Quality Management, etc. are not readily available most of the time. Thus,
productivity improvement can become a costly and difficult exercise. Therefore, there is a
need of a fast and cost effective solution to the problem of productivity improvement.
v
This work offers a time saving productivity improvement and cost optimization
solution for an SME through the implementation of lean methodology and the use of
modern simulation packages (Simio). The implementation of lean methodologies
supports elimination of waste and makes available resources that were incorrectly
allocated, resulting in an increment in productivity. The simulation packages serve as a
visual aid to understand the interrelation between various components of the enterprise.
Additionally, they serve as a tool for the what-if analysis in case of any changes being
made in the existing setup. Most importantly, the use of simulation eliminates the
necessity of multiple trial and error cycles, which in turn saves time and reduces the
overall cost of the improvement effort undertaken by the company. Finally, this research
aims to establish this method of implementation of productivity improvement solution as a
standard for similar SMEs to achieve quick results with minimum cost incurred.
vi
Table of Contents
Acknowledgements .............................................................................................................iii
List of Illustrations .............................................................................................................. ix
List of Tables ....................................................................................................................... x
Chapter 1 Literature review ................................................................................................. 1
Chapter 2 COMPANY AND PRODUCT INFORMATION ................................................... 5
2.1 Company Information ........................................................................................ 5
2.2 Product Information ........................................................................................... 5
Chapter 3 RESEARCH METHODOLOGY .......................................................................... 6
Chapter 4 DIRECT TIME STUDY ....................................................................................... 9
4.1 Data ................................................................................................................... 9
4.2 Analyst’s Notes ................................................................................................ 11
4.3 Calculations ..................................................................................................... 12
4.4 Result ............................................................................................................... 14
Chapter 5 Value Stream Mapping ..................................................................................... 17
5.1 Introduction ...................................................................................................... 17
5.2 Components of VSM ....................................................................................... 18
5.3 Observations .................................................................................................... 21
Chapter 6 Simulation ......................................................................................................... 22
6.1 Problem Definition – ........................................................................................ 22
6.2 Project Planning............................................................................................... 23
6.3 System Definition ............................................................................................. 25
6.4 Conceptual Model ............................................................................................ 28
6.5 Preliminary Design........................................................................................... 32
6.6 Input Data Preparation .................................................................................... 32
6.6.1 Data fitting and test for normality. ............................................................... 33
vii
6.7 Model Translation ............................................................................................ 36
6.8 Verification ....................................................................................................... 37
6.8.1 Machine Utilization ...................................................................................... 38
6.8.2 Operator Utilization ...................................................................................... 39
6.9 Model with Variation ........................................................................................ 40
6.9.1 Machine Utilization ...................................................................................... 41
6.9.2 Operator Utilization ...................................................................................... 42
6.10 Experimentation ............................................................................................... 44
6.10.1 Experiment 1 ........................................................................................... 44
6.10.1.1 Machine Utilization .............................................................................. 45
6.10.1.2 Operator Utilization ............................................................................. 46
6.10.2 Experiment 2 ........................................................................................... 47
6.10.2.1 Machine Utilization .............................................................................. 49
6.10.2.2 Operator Utilization ............................................................................. 50
6.10.3 Experiment 3 ........................................................................................... 51
6.10.3.1 Machine Utilization .............................................................................. 52
6.10.3.2 Operator Utilization ............................................................................. 53
6.11 Analysis and Interpretation .............................................................................. 54
Chapter 7 Conclusion ........................................................................................................ 56
Chapter 8 Future Scope .................................................................................................... 58
8.1 Scope for Improvement ................................................................................... 58
8.2 Scope for Implementation ................................................................................ 59
Appendix A Direct Time Study – Pilot Study ..................................................................... 60
Appendix B Direct Time Study - Sample Observations .................................................... 66
Appendix C Data Fitting - Sample Calculation .................................................................. 69
Appendix D Simio Model Layout ....................................................................................... 72
viii
Biographical Information ................................................................................................... 81
ix
List of Illustrations
Figure 4.1 Pilot Time Study ............................................................................................... 13
Figure 4.2 Direct Time Study – Recorded Data ................................................................ 15
Figure 5.1 Present State Value Stream Map .................................................................... 20
Figure 6.1 Gantt Chart (a) ................................................................................................. 24
Figure 6.2 Gantt Chart (b) ................................................................................................. 24
Figure 6.3 Picture of the facility. ........................................................................................ 28
Figure 6.4 Conceptual Model ............................................................................................ 30
Figure 6.5 Geometric Layout (Courtesy – Autoparts, Inc.) ............................................... 31
Figure 6.6 Model layout- Experiment 2 (Part A) ............................................................... 48
Figure 6.7 Model layout – Experiment 2 (Part B) .............................................................. 48
Figure 6.8 Summary of Throughput .................................................................................. 54
x
List of Tables
Table 4.1 Operation Sequence ......................................................................................... 10
Table 4.2 Allowances for Direct Time Study ..................................................................... 13
Table 4.3 Time Check ....................................................................................................... 16
Table 6.1 Worker allocation in the current system ............................................................ 27
Table 6.2 Conceptual Model ............................................................................................. 29
Table 6.3 Machine Utilization-Deterministic Model ........................................................... 38
Table 6.4 Operator Utilization-Deterministic Model .......................................................... 39
Table 6.5 Stochastic Model Throughput Summary ........................................................... 40
Table 6.6 Machine Utilization – Stochastic Model ............................................................ 41
Table 6.7 Operator Utilization – Stochastic Model ............................................................ 42
Table 6.8 Machine Utilization-Experiment 1 ..................................................................... 45
Table 6.9 Operator Utilization-Experiment 1 ..................................................................... 46
Table 6.10 Machine Utilization-Experiment 2 ................................................................... 49
Table 6.11 Operator Utilization-Experiment 2 ................................................................... 50
Table 6.12 Machine Utilization-Experiment 3 ................................................................... 52
Table 6.13 Operator Utilization-Experiment 3 ................................................................... 53
Table 6.14 Expenses Involved .......................................................................................... 55
1
Chapter 1
LITERATURE REVIEW
According to Prime Faraday Technology Watch, in the UK, SMEs in the
manufacturing sector make up 99% of the firms and provide employment for roughly 50%
of the workforce [10]. However, productivity improvement initiatives have not been very
popular with the manufacturing sector of the economy. Although it has been around for
decades, the lean methodology—and the tools and techniques it offers—are not widely
accepted by the SMEs [1]. Pachanga, Shehab, Roy and Nelder (2006) consider this
reluctance to be the result of the uncertainty that surrounds the cost of implementation
and the potential benefits of the same. Since it is not always possible to foresee the
tangible results of lean implementation, it becomes difficult for the leadership to put their
faith in lean methodologies over conventional methods. Conclusively, the authors state
that a committed leadership is essential for a successful implementation [1].
The paper by Hawkins (2001) highlights the most important tools that SMEs of
this day need in order to maintain their competitiveness and in general, their operational
well being. They include lean manufacturing and supply chain responsiveness, among
other things [10]. To incorporate both attributes, it is imperative to improve the existing
processes. There are multiple barriers that can potentially deter SMEs from implementing
improvements. The barriers include employees that are resistant to change, unavailability
of additional manpower for additional tasks, lack of highly trained specialists, and lack of
standardization [10]. To tackle those barriers, the right question to ask is whether the
SMEs can afford not to improve productivity [10]. Additionally, the right outlook about
productivity improvement is to look at short-term cost (of productivity improvement
initiatives) as long-term investments [10].
2
Many ideas now termed as ‘lean’ stem from the Toyota Production System
(TPS). TPS percolated to Toyota’s suppliers and subsequently entered Japanese
mainstream industry [10]. Value stream mapping (VSM), an important technique offered
by lean methodology, is used for classifying the activities, identifying non-value adding
activities, and eliminating the same for a leaner system performance. Heines and Rich
offer seven tools of VSM which include two new tools, Quality Filter Mapping and
Physical Structure Mapping [11]. Heines and Rich offer a six-step method of VSM
deployment which begins with interviewing the internal stakeholders of the organization,
and the final step being the decision, about which of the seven tools described before are
appropriate for use in a given situation [11]. This novel approach offers a unique way of
implementation of VSM that requires participation from all important stakeholders at all
levels, thereby increasing chances of successful implementation of improvement
initiatives. This method can theoretically be implemented in any industry / facility [11].
Many practitioners have successfully implemented productivity improvement and
waste reduction techniques in the similar niche (i.e. small and medium enterprises).
Gupta and Brannan (1995) implemented Just – In – Time (JIT) in a small manufacturing
company where preliminary analysis identified various problems in the existing setup
[16]. They achieved reduced material movement, reduced lead times, reduced inventory
levels, and a smooth flow of materials [16]. Gunasekaran and Lyu implemented JIT in a
small automotive parts manufacturing facility and observed a continuous annual growth
of 5% to 10% for 3 consecutive years [13]. They noted that improvement in productivity
helps bring the quality defects to surface, in both product and production, which were
invisible prior to the implementation of JIT [13]. Karlsson and Ahlstrom theoretically
implemented the lean principles in an SME manufacturing electronic office supplies and
3
found that the successful implementation of lean is contingent upon lean distribution, lean
procurement, and involvement of partners [14].
A committed leadership with a strong belief in methodologies such as lean, still
needs quantifiable evidence or data to believe in the profitability of the whole exercise of
lean implementation [15]. According to Abdulmalek and Rajgopal, providing quantifiable
evidence in advance can be a difficult task because of the sheer complexity of the
systems [15]. It is nearly impossible to predict the behavior of all the elements in a facility
and about all the Key Performance Indicators (KPIs) because of the dynamic nature and
the vast number of variables involved in a facility [2]. In order to generate belief within the
leadership, the obvious tool to be used is simulation. Simulation offers a capability of
predicting system behavior with preset assumptions. As a result, the combination of lean
and simulation provides a tangible and quantifiable effect analysis of the changes and
improvements proposed at the beginning of the process [2].
In the early phases of process improvement, a variety of opportunities can be
recognized. The prioritization of what opportunities to focus on primarily depends on the
policy set forth by the leadership. Based on the amount of resources available, such as
manpower, funds, and time, the management may choose to tackle a small group of
problems first. In such a case, the combination of lean and simulation tools can be
effective for the impact assessment of the prioritized solutions. Basan, Coccola, and
Mendez (2014) employed a similar strategy while tackling design optimization problems
for a beer packaging line. Various avenues of improvement were identified and
subsequently modeled, and it was up to the leadership to choose which improvements
out of the proposed set of improvements were to be implemented [7].
Simulation can be considered as a proven and reliable tool to check the
requirement, or the lack thereof, of a specific part of the facility. Brown and Sturrock
4
(2009), for a major HVAC manufacturer, deployed the simulation tool (SIMIO) to check
the necessity to structure the assembly line as a belt conveyor, which implies that all the
material movement happens at the same time and at the same pace. They found,
through the simulation results, that there is no absolute necessity for the material
movement to occur as if it was occurring on a belt conveyor and as a result the client
facility plans to remove the belt conveyor and install roller conveyors. This result supports
the utility of simulation tools in driving ‘lean thinking’ as it provides a tangible result that is
essentially what the leadership needs to make a decision [8].
Out of the plethora of simulation tools available in the market, it is necessary to
find the one that best suits the requirements. In broad terms, the objective of this project
was to establish the combination of lean, simulation, and process improvement
techniques as a time-saving, cost-saving way to identify and analyze the changes
necessary for productivity improvement. Accordingly, the simulation tool must possess
the qualities like flexibility, ease of use, ability to perform experiments, etc. C.D. Pegden,
outlines the advantages of SIMIO and observes that it is a graphical, object-oriented tool
which requires comparatively less programming [9]. The software is domain neutral, i.e. it
is designed so that the objects can be configured to act as elements in virtually all
facilities. It also allows creation of new objects if the predefined objects are unable to
meet the requirements of simulation [9]. SIMIO can be used in different paradigms and to
support event, object, process and agent view, and the occurrence of probabilistic events
such as breakdowns, failures, variable lead times, etc. [9]. As it requires less
programming and coding, it does not necessarily require a specifically trained
professional [9]. As it is domain neutral, it can be deployed at any SME without one
having to worry about its applicability [9]. These advantages certainly make SIMIO the
tool of interest for this project [9].
5
Chapter 2
COMPANY AND PRODUCT INFORMATION
2.1 Company Information
To protect the anonymity of the company, this facility is called ‘Autoparts Inc.’ for
the purpose of this project. Autoparts Inc. is a leader in manufacturing sheet metal press
components in Maharashtra, India. Established in 1967 with small and limited facilities,
the company now possesses diverse production facilities to meet varying customer
needs and schedules.
Autoparts Inc. has nurtured development, production, and marketing functions to
ensure desired quality, reliability, reasonable price, and timely delivery to create values
for customer businesses. The group supplies sheet metal press components to Indian
and overseas OEMs with consistent growth in intricacy, range, and values. Their
business philosophy is to create customer end values for continuous, sustained, and
mutual growth.
2.2 Product Information
This research focuses on Autoparts Inc., specifically one division of their
manufacturing known as unit 3. The products manufactured by Autoparts Inc. industries
at Unit 3 are the sheet metal discs required to cover the brake shoes of TATA ace and
Piaggio Vehicle Pvt. Ltd. To manufacture the product, multistage forming processes are
carried out. The processes consist of shearing, blanking, forming, vertical drilling, inclined
drilling, deburring and visual inspection. The processes are carried out on various
presses and other machines on the shop floor. The presses are equipped with the
required dies for the forming and drilling operations.
6
Chapter 3
RESEARCH METHODOLOGY
The first step of the project was data collection. For successful implementation of
the methodology, a large amount of data is needed to ensure high accuracy of the
results. This data was acquired from an automobile spare parts manufacturing facility in
India. The data to be collected was primarily the work measurement data, along with
geometric layouts, production statistics, operator details, actual demand, demand
forecast, etc. A sample dataset was collected at first. Based on the sample dataset, the
number of readings required for the required confidence level and accuracy was
calculated.
After data collection was complete and the required number of data points were
obtained, the next step was to perform a time study. A time study is a method of
calculating the standard time of production for an average component, produced in a
predefined environment. Time study considers the worker performance rating and the
PFD (Personal, Fatigue, Delays) allowances. As a result, a standard time for the product
was obtained which is indicative of the ideal time it takes to manufacture one component.
After the time study was complete, we moved forward to the next stage (i.e.
application of lean). In this stage, firstly Value Stream Mapping (VSM) was deployed.
This tool is useful in identifying value adding, non-value adding, and necessary non-value
adding activities. Secondly, the necessity and possibility of application of 5-S was
checked. The inventory levels and transporter routing logic was assessed to identify the
waste of unnecessary inventory and waste of transportation and excess motion.
After the opportunities of waste reduction were identified, the next step was to
understand and learn the simulation package chosen for use during this project. The
package chosen was SIMIO as it offers several considerable advantages. This package
7
requires less coding and programming as compared to other packages available off the
shelf. It is graphical and can process large amounts of data with relative ease. These
characteristics make it lucrative, as the objective of this project is to establish this method
of productivity improvement and cost optimization as a standard for manufacturing
oriented SMEs. The target audience in the industry may not have an adequate amount of
specially trained workforce available thus; the ease of use offered by SIMIO was a key
factor in its selection.
After necessary prowess was achieved in the use of SIMIO, the modeling phase
commenced. In the first phase of modeling, the factory floor was modeled using the
deterministic / average values. This model should ideally produce results closer to the
average performance of the factory. A certain level of error is permitted as a deterministic
model is bound to be subject to some approximation. The acceptable rate of error is set
at 5%. This means that the model with calculated throughput within +/- 5% of the actual
value will be accepted. Successful completion of this phase indicates that the data and
the modeling assumptions are representative of the real-world situation.
In the next stage, the complexity of the model was increased. The model was
now required to possess probability distributions such as those observed in the obtained
dataset in stage 2. This model should ideally give different throughput numbers every
run, since the probabilistic approach is now incorporated in the model. However, the
average throughput over several production runs should be similar to that of the actual
production of the facility. The acceptable rate of error at this phase is set at 3%.
The next step was to brainstorm the opportunities for improvement. Possible
improvements can be the solutions to the individual problems identified in the fourth
stage or some combination of them. These solutions will then be incorporated into the
model and the effect of the changes on KPIs such as throughput levels, inventory levels,
8
worker utilization will be observed. The selection of the best solution depends on the
priorities set forth by the company leadership.
The next and final stage was to document and record the findings and
observations of this project that can serve as a template for future use by manufacturing
oriented SMEs. The advantages of following this method of productivity improvement, as
described in the project, was emphasized upon so that the other SMEs can implement
this method for their benefit.
9
Chapter 4
DIRECT TIME STUDY
4.1 Data
To complete the direct time study, a large amount of data was needed. The data
was collected at the factory.
In this time study analysis, the following method of data collection was followed.
Firstly, measurable work elements were defined. Then, the time taken by the workers to
complete these work elements was observed. Subsequently, worker’s performance rating
was assigned. The measurable work elements found were as follows.
• Shearing Operation – Setup Time
• Shearing Operation – Process Time
• Blanking Operation – Setup Time
• Blanking Operation – Process Time
• First Forming Operation – Setup Time
• First Forming Operation – Process Time
• Second Forming Operation – Setup Time
• Second Forming Operation – Process Time
• Stamping Operation – Setup Time
• Stamping Operation – Process Time
• Cutting and Bending Operation – Setup Time
• Cutting and Bending Operation – Process Time
• Hole Pressing Operation – Setup Time
• Hole Pressing Operation – Process Time
• Tube Hole Operation – Setup Time
10
• Tube Hole Operation – Process Time
• Contouring Operation – Setup Time
• Contouring Operation – Process Time
• Grinding Operation
• Packing Operation
Time taken to accomplish each of these activities was recorded using the
standard time study procedures. The method of recording used was the Snapback
Method. The rating was generally assumed at 90%.
The operation sequence is shown in the table below.
Table 4.1 Operation Sequence
Workstation No Machine Present Process Present Operational
sequence
1 Press Blanking 2
2 Press
3 Press 2nd forming/ drawing 4
4 Press 1st forming/ drawing 3
5 Press Cutting/ Bending 6
6 Drill Drill Vertical 7
7 Drill Drill Inclined 8
8 Deburring Deburring 9
9 Inspection Table Inspection 10
10 Press
11 Press
11
Table 4.1 – Continued
12 Press
13 Press
14 Press Stamping 5
15 Press
16 Shearing M/c Shearing 1
17 Hand Grinders Grinding 11
18 Packing area Packing 12
4.2 Analyst’s Notes
The number of occurrence for every work element was not the same. One cycle
of blanking operation created an equivalent 4 parts for the process. Therefore, the
number of occurrence was set at 0.25. One cycle of shearing operation created
equivalent of 40 parts. Therefore, the number of occurrence was set at 0.025. One cycle
of packing operation processed 25 parts at the same time. Therefore, the number of
occurrence was set at 0.04. Since the observations were recorded over a period of
several days, the TEBS and TEAF don’t have a single value. Alternatively, a cumulative
value is provided. The start time and the end time don’t have a single value but the total
elapsed time was noted carefully. Observations, wherein the time taken was
disproportionately large due to some failure in the machinery, were carefully noted and
discarded. There are some outliers in the time study. These can be attributed mostly to
mishandling of the equipment and the jobs. For example, there were cases when the job
would slide out of the operator’s grasp and the operator would have to pick it up.
12
4.3 Calculations
Number of cycles required –
The formula used to calculate the number of cycles (Nc) required is as follows.
Nc = ((tα/2 * s)/(k*μ))^2
Where,
tα/2 = t value for (α = 0.05) i.e. confidence interval of 95%
s = Standard deviation of the sample observations
K = Accuracy or proportion of interval (accuracy = +/- 5% of the actual)
μ = sample mean
Per this formula, a sample of 15 readings was taken for each process element.
The sample calculations of the number of readings required per the dispersion of values
can be found in the table below. The full spreadsheet of work measurement data
recorded can be found is Appendix A.
13
Figure 4.1 Pilot Time Study
After calculating the number of observations required for required accuracy and
confidence interval, the highest number was found out to be 147 for Contouring – Setup
time. Hence the required number of observations was selected as 150. Standard
procedure for calculations was used with allowance shown in the table below.
Table 4.2 Allowances for Direct Time Study
Category Value
Personal 5%
Basic Fatigue 4%
Variable Allowance – Abnormal Position
Allowance
2%
Variable Allowance – Use of Force 1%
14
Table 4.2 – Continued
Variable Allowance – Noise Level 2%
Variable Allowance - Monotony 1%
Total 15%
According to the International Labor Organization rules, the allowances can be
explained as follows [17]. A personal allowance of 5% is necessary for all operations
along with a basic fatigue allowance of 4%. Additionally, an abnormal position allowance
of 2% is awarded since the working positions are slightly awkward requiring long periods
of bending. A 1% allowance is awarded for use of force since the job involves lifting up to
10 lbs. Noise level allowance of 2% is awarded because of the presence of intermittent
noises and 1% allowance is given for monotony. Cumulatively, a total of 15% allowance
is awarded.
The recorded work measurement data can be found in Figure 4.2, shown below.
4.4 Result
After Using the formula,
Standard Time = Normal Time * (1+ Allowances), the standard time, after all due
consideration, was found out to be 1.15 minutes.
The input required to calculate error is the total recorded time and the
unaccounted time, the values of which were 204.27 and 3.08 minutes respectively. The
error after calculations was found out to be 1.48% that is within the acceptable limit (0%
to 2%).
16
The time check can be shown as follows.
Table 4.3 Time Check
Time Check
Finishing Time (minutes)
Starting Time (minutes)
Elapsed Time (minutes) 714.5
TEBS (minutes) 0.48
TEAF (minutes) 0.13
Total Check time (minutes) 0.61
Effective Time (minutes) 703.35
Ineffective Time (minutes) 0
Total Recorded Time (minutes) 703.956
Unaccounted Time (minutes) 10.54405
Recording Error 1.48%
17
Chapter 5
Value Stream Mapping
5.1 Introduction
Value Stream Mapping (VSM) is the conventional method of charting the process
under observation into distinguishable parts. These processes can be categorized into
three distinct types: Value - adding processes, Non - value adding processes, Necessary
non - value adding processes. An explanation of what these categories entail, along with
some examples can be found below.
• Value - adding processes – These processes involve operations where work is
being done on the component / product through direct or indirect manual labor,
with or without the use of machines [11]. The typical examples of value – adding
processes include cutting, drilling, assembly, forging, etc.
• Non - value adding processes – These processes involve operations or actions
that are entirely unnecessary and should be eliminated to the highest extent
possible. These processes are the most likely causes of waste in the system
[11]. The typical examples include waiting time, stacking, work in process (WIP)
inventory, etc.
• Necessary non – value adding processes, as the name suggests, these
processes do not typically add value to the product but are necessary under
current manufacturing / operating environment [11]. To eliminate the waste
introduced in the system due to these processes, a redesign of the
manufacturing / operating environment is necessary. The examples include
material movement, packing and unpacking tools, worker movements, etc.
18
5.2 Components of VSM
While preparing the current state value stream map for Autoparts Inc., the
following components of the manufacturing environment were taken into consideration.
• Supplier - This entity supplies the main raw material to the facility. The main raw
material is sheet metal that is supplied on a fixed daily basis. The transport is
facilitated through trucks.
• Customer - Finished parts are shipped to this entity after packaging. The
transport is facilitated through trucks.
• Process control - This entity is the representation of the management of the
facility. This entity provides inputs in form of orders to the supplier and receives
inputs in form of orders from the customers. This entity also provides inputs to
the first station in the process regarding the quantity and the commencement of
processing. It receives input from the final station regarding the completion of
process.
• Workstation - This entity represents any active workstation in the process where
an operation takes place on the component. A workstation may have a specific
number of operators assigned to it.
• Operator – This entity represents the operators working on the shop floor. Some
operators are assigned to specific workstations whereas others are shared
between multiple workstations or processes.
• Inventory – This entity represents the inventory between two workstations. On
the shop floor, material movement is carried out in stacks of 40, thereby
accumulating an inventory of 80 parts in between workstations.
19
• Timeline – Timeline is used to denote the time it takes to complete a specific
activity. The crest part of the timeline denotes a Non-Value adding activity
whereas a trough part denotes a Value- adding activity.
• Activity ratio – The activity ratio is calculated by dividing the value-added time by
the production lead time. The value-added time is the sum of times during which
value is added, whereas the production lead time is the sum of non-value added
time.
• Information flow – Different types of arrows are used to show the direction and
the mode of the flow of information. A straight arrow means manual / handwritten
flow of information whereas the lightening arrow means electronic flow of
information.
The resulting VSM diagram can be found in below.
21
5.3 Observations
Following observations can be made based on the value stream map.
• There is a high inventory level in between workstations.
• The activity ratio is very poor, i.e. 1%.
• The Non-value adding activities can be reduced to a great extent.
22
Chapter 6
Simulation
After completion of value stream mapping, the next step in this project is
simulation. For a structured effort and standardization, simulation projects are conducted
in these sequential steps. They are as follows.
6.1 Problem Definition –
Setting a problem definition involves defining the areas or the aspects of the
plant that the project seeks to improve. It also intends to answer the ‘why’ of the project.
Having a clear problem definition will help to have a clear set of objectives and a minimal
obscurity in the project goals. In this study, we shall try to tackle an industrial situation
where, a managerial team must take decisions about expansion of their sheet metal
pressing plant. We have collected the data for current manufacturing conditions at this
plant. We have studied their charts to know the manufacturing process and its
constraints. Their current production ranges somewhere between 1800 - 1850
components per shift. They have an increase in demand that requires them to increase
their production almost immediately. It is predicted that this increase in demand is not a
momentary flair, but will continue to rise. It will require a significant capacity increase. It is
also mandatory that the suggested changes should not require a substantial capital
investment and that the changes should not increase the rigidity of the process. The
management thinks that the plant needs to have flexibility in the products manufactured.
Currently, the manufacturing line at Autoparts Inc. is an imbalanced system in terms of
operator utilization and machine utilization. Also, the overall output is far below the
system potential according to the management. Therefore, our goal is to balance the
system and to increase the system output at minimum possible expense. We want to
23
provide a turn-key solution for the productivity improvement problem which can be used
and reused by moderately skilled professionals in a similar environment, any time in the
future.
Problems faced by the company –
Currently, the plant is struggling to match the demand. Problems that the company is
facing now are,
• Production output is currently below market demand.
• The machines and operators are underutilized.
• The plant is not efficient in terms of resource utilization.
The constraints under which the analysts are required to operate are as follows.
• Avoid large investments in presses.
• Maximize the utilization of the available resources.
• Avoid layoffs as much as possible.
• Provide a long-term solution for flexibility.
6.2 Project Planning
This step involves planning and managing the resources necessary for the
duration of the project.
• Time – The management of the time required for each step can be found in the
Gantt Chart shown in Figure 6.1 and Figure 6.2. Postscript ‘I’ in the Gantt Chart
refers to the first of a month and postscript ‘II’ refers to the second half.
• Manpower – This task will be performed by the analyst with the help and
guidance of the Supervising Professor.
24
• Equipment – The equipment required for this project is a stopwatch and a
personal computer to operate the DES tool Simio.
• Software – We have chosen Simio as the discrete event simulation tool. The
reasons for choosing simio are as follows.
▪ Less programming required as compared to other simulation tools.
▪ Excellent Graphical User Interface (GUI).
▪ Domain Neutral nature.
Figure 6.1 Gantt Chart (a)
Figure 6.2 Gantt Chart (b)
25
6.3 System Definition
To achieve the objectives of the project, it is highly important to define the system
that we intend to work on. The system definition phase required us to understand the
system that we wish to simulate during the project. The system can be described as
follows.
• The main raw material that is purchased from the vendors is steel sheets, which
arrives in stacks of 80 sheets per day. The sheets are first cut in 10 parts of
rectangular shape using the shearing machine, which needs two operators.
• These rectangular cut parts are stacked in the output buffer of the shearing
machine. A helper transports these parts with the help of a trolley at the rate of
40 cut pieces per trip to the blanking machine, which is the subsequent
operation.
• The parts are stacked up in the input buffer of the blanking machine. The
blanking machine, which needs one operator, cuts these parts into circular discs.
It cuts 4 circular discs from each rectangular cut piece. Thus, each sheet
effectively gives 40 discs.
• The circular discs are stacked up in the output buffer of the blanking machine.
From this buffer, the parts are transported to the input buffer of Press 3 where 1st
forming/drawing is done. This machine requires one operator for processes and
one operator for material movement.
• After press 3 performs its functions, the parts are stacked up in the output buffer
from where they are transported to the input buffer of Press 2. Press 2 performs
2nd drawing and the parts are again stacked up in its output buffer. This machine
requires one operator.
26
• From the output buffer of Press 2, the parts go to the input buffer of Press 7
where stamping is done. This machine requires one operator to process the
components and one operator to simply transport the materials to and from this
machine.
• After Press 7, parts are brought to the input buffer of Press 4, for the cutting and
bending operation. Press 4 also requires one operator.
• After the cutting and bending operation, the parts are stacked up in the output
buffer. From here, they are taken to the input buffer of the Hole Pressing
operation, which requires one operator.
• After Hole Pressing is done, the parts, after being stacked up in the output buffer.
They are then transported to the input buffer of the Tube Hole Operation, which
requires one operator.
• After the Tube Hole operation is carried out, the parts are stored in the output
buffer, ready to be transported to the Deburring machine. Deburring machine
needs one operator.
• After deburring is done, these parts are stacked up in the output buffer. From
here, they will be taken to the grinding section.
• The input buffer of the grinding section serves as a buffer for the grinding
operation part input. There are currently 3 operators who perform this operation
using hand held grinders.
• The subsequent operation is packaging operation which requires 2 operators.
Here, the finished parts are bundled in batches of 25 and packed for shipment.
The worker allocation is shown in table below.
27
Table 6.1 Worker allocation in the current system
Workstation No of Operators
Shearing 2
Blanking 1
First forming 1
Second forming 1
Stamping 1
Cutting & Bending 1
Hole pressing 1
Tube Hole 1
Contouring 1
Grinding (3 Workstations) 3
Packaging 2
Helpers 4
Total 19
Additionally, there is a supervisor and a plant engineer present during the shift.
The actual system is shown in figure below.
28
Figure 6.3 Picture of the facility.
6.4 Conceptual Model
The conceptual model can be termed as a model that reflects all the vital process
parameters as they appear on the actual factory floor. In the case of this project, the
conceptual model should contain the following elements.
• Machines
• Workers
• Material handling
• Source of materials
• Flow of material
29
These elements that are present on the factory floor were suitably termed as per the
convenience of the software. Below is the list of the conventional terms and their
respective software specific translation.
Table 6.2 Conceptual Model
Conventional Terms Software Specific Translations
Presses Workstations
Operator Worker
Material handling Vehicle
Source of materials Source
Stacks Queue
Shearing Machine Separator
Packaging Combiner
The geometric layout of the factory is shown in the figure below. It contains two parallel
rows of machines placed at a distance from each other. In between the two rows is the
storage place where earlier setup items are stored. The middle row also contains the
shearing setup. The material receipt and dispatch is accomplished through the door at
the eastern side of the plant. The entry and exit of the personnel is facilitated through the
door at the southern side of the plant. The actual layout was provided by the company,
but the remaining details, however trivial, were measured later to achieve maximum
accuracy of the geometric layout.
32
6.5 Preliminary Design
After the conceptual model is ready, the system performance measures should
be selected. The performance measures indicate the factors that are important or
decisive in the net performance of the company. Having the detailed plan at this early
stage facilitates better understanding of the system and the system variables.
The process variables that we have chosen to vary are as follows.
• System throughput – System throughput is of the prime concern and the project
will focus on maximizing the system output.
• Machine utilization – It is imperative that the machines stay in the processing
mode for an optimum amount of time. The ideal machine utilization should be
between 60% to 80%.
• Operator utilization – The operators should be busy for an optimum amount of
time.
• Material movement – Material movement should not be a reason for the
sluggishness in the production. We will focus on improvising the existing modes
of material movement. We will try to decrease the amount of material movement
by adding some new equipment, if necessary.
6.6 Input Data Preparation
For this project, the data recorded during the direct time study was used.
Additional data analysis was performed to identify the probability distributions in the data.
The work measurement observations can be found in Appendix A and Appendix B.
33
6.6.1 Data fitting and test for normality.
After recording the work measurement data, it is necessary to identify the
probability distributions of the recorded data to understand the patterns. Usually, the most
common distribution observed in natural events is the normal distribution, and hence it
will be a priority to check for normality in the datasets. If normality can’t be confirmed
using the statistical analysis tools, we move on to the next possible common data
distributions e.g. exponential, lognormal, gamma, beta, etc. If the data does not
accurately conform to the requirements of the probability distributions listed, we can
assume a triangular distribution, which can be arbitrarily observed in the data. The DES
tool chosen for this project (Simio) supports most probability distributions. By
incorporating the behaviors of these probability distributions, the simulation model can
mimic the real-world situation more accurately.
The stepwise method employed for data fitting of a particular dataset is as
follows. This process is performed on each dataset recorded during the direct time study.
• For a dataset, generate a QQ plot using R and a Histogram using MS-Excel. QQ
plots and histograms are fairly accurate tools to check for normality in the
dataset.
• Conduct the Shapiro -Wilk test using R. Comparing the p-value obtained from the
results of the Shapiro-Wilk test with the predetermined level (0.05) we can judge
whether there is a confirmed normality in a given dataset.
• If both these steps lead us to a result of normality not being observed in the
dataset, we move on to the next step i.e. Chi Square test for goodness of fit. This
test is explained below.
• The Chi-Square test is used to test if a sample of data belongs to a population
that has a behavior resembling that of a specific probability distribution. This test
34
can be used for discrete data that is divided in classes. It also requires a
significant number of samples for it to be effective and valid. Since, both these
conditions are satisfied by the data that we have, the Chi-Square test can be
used here. The hypothesis for this test is as follows
H0: The data follows a specific distribution.
H1: The data does not follow a specific distribution.
Test condition: Reject H0 if χ 2 > χ 2 (1-α, k-c)
Where,
α = Significance level (0.05)
k = Number of classes/bins (15)
c = Number of estimated parameters for a specific distribution +1
Examples,
▪ Normal distribution –
For the normal distribution, two parameters (mean and standard deviation)
are required. Hence, c=2+1=3
Therefore,
The hypotheses are
H0: The data follows a Normal distribution.
H1: The data does not follow Normal distribution.
Test condition: Reject H0 if χ 2 > χ 2 (1-0.05, 15-3)
Similarly,
▪ Exponential distribution –
H0: The data follows Exponential distribution.
35
H1: The data does not follow Exponential distribution.
Test condition: Reject H0 if χ 2 > χ 2 (1-0.05, 15-3)
▪ Poisson distribution
H0: The data follows a Poisson distribution.
H1: The data does not follow Poisson distribution.
Test condition: Reject H0 if χ 2 > χ 2 (1-0.05, 15-3)
▪ Gamma distribution
H0: The data follows a Gamma distribution.
H1: The data does not follow Gamma distribution.
Test condition: Reject H0 if χ 2 > χ 2 (1-0.05, 15-3)
▪ Lognormal
H0: The data follows a Lognormal distribution.
H1: The data does not follow Lognormal distribution.
Test condition: Reject H0 if χ 2 > χ 2 (1-0.05, 15-3)
According to the Chi-Square distribution table, for the same level of
significance, a smaller Chi-Square value is associated with lesser degrees of
freedom. Hence, in order to avoid unnecessary calculation, we will test the
hypothesis at c=4. If in any case, we fail to reject the hypothesis, we can go
into the additional analysis necessary to identify which distribution the data
conforms with.
The sample results of data analysis can be found in Appendix C.
36
6.7 Model Translation
The actual operations being carried out on the floor need to be converted into
equivalent and suitable objects so that they can be used in the software i.e. Simio. Thus,
we need to translate the model into a suitable format.
The translation of the model can be explained as below.
▪ Shearing machine - Shearing machine, as the name suggests, cuts a sheet into
10 equal rectangular pieces of the designed dimensions. As the machine
separates one part into several parts, it is equivalent to a separator according to
the Simio logic.
▪ Blanking machine - The blanking machine chiefly serves the purpose of punching
the cut sheets, to produce 4 circular discs from each of the pieces. As it
separates one component into several components, it can be termed as a
separator.
▪ Presses – The presses used in the plant are required to carry out the various
functions. They can be termed as workstations. The workstations under this
category are
▪ First Forming
▪ Second Forming
▪ Stamping
▪ Cutting and bending
▪ Hole Pressing
▪ Tube Hole
▪ Contouring
37
▪ Grinding Machines – The grinding machines in this facility are simple hand held
grinders. There is no setup time associated with this operation and hence for the
ease of programming, we use server objects to imitate the 3 grinding machines.
▪ Packaging – At the Packaging station, finished parts are packed in stacks of 25.
To resemble this operation, combiners are used.
▪ Queues – Queues primarily achieve the objective of storing the materials. A
queue is required to imitate input buffers, output buffers, and processing buffers.
▪ Operators – Operators employed at their respective stations perform their
respective tasks. Additionally, there are helpers which help in various tasks and
assist in material movement. The worker is a versatile resource. It can be used
as a secondary resource or a vehicle class object in Simio.
6.8 Verification
To Verify the validity of the translated model, we build a deterministic model
which considers only the average/ standard time of every process. Here, no variation is
considered whatsoever.
After running the model, we find that the throughput of the simulated process for
one shift of 8 hours is 1750 finished components ready for dispatch. This is very close to
the real-world system. The error in this model is 3%. Therefore, it can be said that the
model generated so far is representative of the actual system and ready for further
modifications. The layout of the model is shown in Appendix D.
The assumptions are as follows,
• The skill expertise of the operators is same for all operations.
• Machine failures are not considered.
• Buffer capacity is infinite.
38
• Movement speed of the vehicle and operator is same for all tasks
• The sheets arrive at once, every day.
The KPIs observed in this model are as follows.
6.8.1 Machine Utilization
Table 6.3 Machine Utilization-Deterministic Model
Machine Utilization (%)
Shearing 8.32
Blanking 25.39
First Forming 62.82
Second Forming 87.85
Stamping 39.40
Cutting and bending 71.73
Hole Pressing 73.06
Tube Hole 58.15
Contouring 47.14
Grinder 1 9.78
Grinder 2 10.88
39
Table 6.3 – Continued
Grinder 3 13.81
Packaging 22.63
6.8.2 Operator Utilization
Table 6.4 Operator Utilization-Deterministic Model
Operator Utilization (%)
OP_SH_1 22
OP_SH_2 22
OP_BL_1 25.39
OP_FF_1 62.82
OP_FF_2 14.31
OP_SF_1 87.85
OP_ST_1 39.4
OP_ST_TRANSPORT 18.52
OP_CB_1 71.73
OP_HP_1 73.06
OP_TH_1 58.15
OP_CT_1 47.14
40
Table 6.4 – Continued
OP_GR_1 9.78
OP_GR_2 10.88
OP_GR_3 13.81
OP_PK_1 22.6
OP_PK_2 22.6
VEHICLE1 90
VEHICLE2 2
6.9 Model with Variation
To impart variability into the model, which is always present in the real-world
system, we use the probability distributions that we identified in the input data preparation
step. The addition of probability distribution means that the time required for each
operation will now be randomly selected from the assigned probability distribution tables.
Resulting system output will vary each time, like that of the real-world production system.
After 20 simulated runs of the model, the throughput was as follows
Table 6.5 Stochastic Model Throughput Summary
Run Throughput Run Throughput Run Throughput Run Throughput
1 1850 6 1857 11 1800 16 1846
41
Table 6.5 – Continued
2 1828 7 1820 12 1840 17 1848
3 1825 8 1827 13 1844 18 1835
4 1833 9 1852 14 1831 19 1846
5 1830 10 1838 15 1835 20 1839
The average throughput of the 20 runs was found out to be 1840
This throughput is close to the actual throughput with an error of less than 1%. Some of
the key observations are listed here.
1. The flow of material is very complex
2. The amount of time for which the machines and operators are utilized is shown in
the tables below.
6.9.1 Machine Utilization
Table 6.6 Machine Utilization – Stochastic Model
Machine Utilization (%)
Shearing 8.79
Blanking 27.5
First Forming 65
Second Forming 75
Stamping 45
42
Table 6.6 – Continued
Cutting and bending 80
Hole Pressing 70
Tube Hole 55
Contouring 55.21
Grinder 1 10.5
Grinder 2 10.5
Grinder 3 8.75
Packaging 27.5
6.9.2 Operator Utilization
Table 6.7 Operator Utilization – Stochastic Model
Operator Utilization (%)
OP_SH_1 23
OP_SH_2 23
OP_BL_1 27.5
OP_FF_1 65
OP_FF_2 14.31
OP_SF_1 75
43
Table 6.7 – Continued
OP_ST_1 45
OP_ST_TRANSPORT 18
OP_CB_1 80
OP_HP_1 70
OP_TH_1 55
OP_CT_1 55.21
OP_GR_1 10.5
OP_GR_2 10.5
OP_GR_3 8.75
OP_PK_1 27.5
OP_PK_2 27.5
VEHICLE1 90.42
VEHICLE2 2
It can be seen that the packaging machine’s operators are idle for most of the
time. They could be used further down the line. Shearing and blanking machines are also
starved for parts.
44
6.10 Experimentation
Now that we have accurately modeled the existing setup of Autoparts Inc., we
can try to implement various changes to the simulated model to understand their impact
on the system behavior. We use the waste identification and reduction techniques
outlined in the lean methodology. According to the lean methodology, there are seven
types of waste that can exist in a system. They are:
• Transport – Waste of transport
• Inventory – Waste of components in the inventory that are not being processed
• Motion – Waste of unnecessary motion that the objects or the operators make
• Waiting – Waste of objects waiting for the subsequent operation
• Overproduction – Waste of excess production
• Overprocessing – Waste of excess processing resulting from poor setup
• Defects – waste of identifying or reworking on defects.
We shall try to find the presence of these defects and try to minimize them as
much as possible.
6.10.1 Experiment 1
In this experiment, we shall strive to minimize and simplify the material
movement on the shop floor. The material from blanking goes to press 3, then comes
back to Press 2, then across the shop floor to Press 7 and then back to Press 4. This is
adding a massive bottleneck due to the material movement necessary. We propose that,
the dies and punches should be rearranged to regulate the material flow in a straight line.
This way, the operators will spend more time in processing and less time in travelling to
and from a workstation. The modified layout can be shown as below. Additionally, the
45
quantity of the batch of materials, which is currently set at 40, should be reduced to 30.
This way there will be less WIP inventory in the system.
After making the necessary changes and running the model for 20 simulated
runs, the average throughput is 2000 components in one shift. The improvement in
throughput is 9%.
The KPIs are found to be as follows
6.10.1.1 Machine Utilization
Table 6.8 Machine Utilization-Experiment 1
Machine Utilization (%)
Shearing 8.35
Blanking 25.44
First Forming 54.56
Second Forming 77.03
Stamping 34.56
Cutting and bending 62.55
Hole Pressing 64.15
Tube Hole 55.32
Contouring 47.96
Grinder 1 16.95
46
Table 6.8 – Continued
Grinder 2 12
Grinder 3 9
Packaging 28
6.10.1.2 Operator Utilization
Table 6.9 Operator Utilization-Experiment 1
Operator Utilization (%)
OP_SH_1 24
OP_SH_2 24
OP_BL_1 25.44
OP_FF_1 54.56
OP_FF_2 15.58
OP_SF_1 77.3
OP_ST_1 34.56
OP_ST_TRANSPORT 20.04
OP_CB_1 62.55
OP_HP_1 64.15
OP_TH_1 55.32
47
Table 6.9 – Continued
OP_CT_1 47.96
OP_GR_1 17
OP_GR_2 13
OP_GR_3 9
OP_PK_1 28
OP_PK_2 28
VEHICLE1 93
VEHICLE2 3
The Key Observations here is that the transporter object is busy for 93% of the time.
6.10.2 Experiment 2
The transporter being busy for 93% of the time is a serious cause of concern. To
alleviate this bottleneck, we introduce a conveyor system for the material movement
between the two sets of workstations that have the highest distances between them.
They are:
• From Shearing Machine to Blanking Machine – 6m.
• From Contouring Machine to the Grinders – 7m.
After making the necessary change, the layout is as follows.
49
After making the changes and running the model, we find that the average
throughput of the model is 2250. This goes to prove that the material movement was a
bottleneck in the system. The KPIs are as follows.
6.10.2.1 Machine Utilization
Table 6.10 Machine Utilization-Experiment 2
Machine Utilization (%)
Shearing 8.65
Blanking 26.35
First Forming 62.43
Second Forming 86.91
Stamping 39.67
Cutting and bending 71.38
Hole Pressing 73.03
Tube Hole 63.32
Contouring 53.48
Grinder 1 13.43
Grinder 2 15.36
Grinder 3 13.65
Packaging 29.21
50
6.10.2.2 Operator Utilization
Table 6.11 Operator Utilization-Experiment 2
Operator Utilization (%)
OP_SH_1 25
OP_SH_2 25
OP_BL_1 26.35
OP_FF_1 62.43
OP_FF_2 17.92
OP_SF_1 86.91
OP_ST_1 39.67
OP_ST_TRANSPORT 22.91
OP_CB_1 71.38
OP_HP_1 73.03
OP_TH_1 63.32
OP_CT_1 53.48
OP_GR_1 13.43
OP_GR_2 15.36
OP_GR_3 13.65
51
Table 6.11 – Continued
OP_PK_1 29.21
OP_PK_2 29.21
VEHICLE1 85
VEHICLE2 3
6.10.3 Experiment 3
We can observe from the results of experiment 2 that the grinding machines are
busy only for 12% to 13% of the times. Additionally, due to the simplified material
movement, the stamping machine transport operator is busy only for 22% of the time.
Also, the number of entities exiting the grinding stations was 2353 but only 2250 of them
were packaged subsequently. Based on these observations, we propose the following
changes.
1. Remove Grinder 3 – The operator will be available for other assignments.
2. Allocate the stamping machine transportation to another shared transporter –
The stamping machine transport operator will be available for other assignments.
We allocate these operators to an additional packaging station. In this facility, the
task of packaging is done simply by making stacks of 25 finished components and
securing them together. Making these changes to the model and running it, we get an
average throughput of 2375 finished components per shift. The increment in productivity
is 5.5%. The KPIs are as follows.
52
6.10.3.1 Machine Utilization
Table 6.12 Machine Utilization-Experiment 3
Machine Utilization (%)
Shearing 8.47
Blanking 26.44
First Forming 62.66
Second Forming 88.56
Stamping 39.46
Cutting and bending 71.35
Hole Pressing 72.82
Tube Hole 63.70
Contouring 54.85
Grinder 1 21.56
Grinder 2 21.86
Packaging 1 15.25
Packaging 2 15.65
53
6.10.3.2 Operator Utilization
Table 6.13 Operator Utilization-Experiment 3
Operator Utilization (%)
OP_SH_1 25
OP_SH_2 25
OP_BL_1 26.44
OP_FF_1 62.66
OP_FF_2 17.82
OP_SF_1 88.56
OP_ST_1 39.46
OP_ST_TRANSPORT 22.92
OP_CB_1 71.35
OP_HP_1 72.87
OP_TH_1 63.7
OP_CT_1 54.85
OP_GR_1 21.56
OP_GR_2 21.86
OP_GR_3 (at packaging 2) 15.65
54
Table 6.13 – Continued
OP_PK_1 15.25
OP_PK_2 29.21
VEHICLE1 91
VEHICLE2 3
6.11 Analysis and Interpretation
The simulation initiative has been successful in identifying the opportunities of
improvement and studying the impact various changes on the overall system behavior.
The summary of results of the various stages of simulation can be shown in a bar chart
as follows.
Figure 6.8 Summary of Throughput
55
As per the goals set at the initial stage, we have taken the system through the
entire simulation process to increase the productivity of the system at minimum
expenditure. The total improvement in productivity earned over these experimentations
can be given by,
Total Increase in Productivity = 𝐹𝑖𝑛𝑎𝑙 𝑇ℎ𝑟𝑜𝑢𝑔ℎ𝑝𝑢𝑡−𝑂𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑇ℎ𝑟𝑜𝑔ℎ𝑔𝑝𝑢𝑡
𝑂𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑇ℎ𝑟𝑜𝑢𝑔ℎ𝑝𝑢𝑡
= (2375 - 1800) / (1800)
= 31%
The expenses that will be incurred are shown below. This report will be submitted
to the industry for perusal. It will be their decision whether to opt for these modifications.
Table 6.14 Expenses Involved
Experiment Additional Investment Required
Experiment 1 No investment required
Experiment 2 Conveyor
36’ @ INR10000/running foot
= INR360000 ($5200)
Experiment 3 No investment required
As we can see, the proposed changes require minimal capital investments. The
expenses involved in rearranging the dies and installation of conveyors will be assessed
and taken care of by the management, if they choose to implement the changes.
56
Chapter 7
Conclusion
It is motivating to understand that this method has positive results. We set out
with some constraints that we were supposed to operate under. These constraints were
set forth by the management of the facility. Comparison of the final result with the original
constraints and problem statement will help us assess the effectiveness of the proposed
changes.
The first constraint dictated that investments in presses were to be avoided. This
was partly because of the fact that there were additional presses present with the facility
and they can be repurposed to carry out any pressing operation with changes in dies and
fixtures. This constraint is strictly followed throughout the course of the project. None of
the experiments require the purchase of a press. Experiment 1 however requires the
press 2 to be returned to working conditions from its current nonfunctioning state. This
will require minimal investment and no purchases whatsoever.
The second constraint required us to maximize the utilization of the resources
already present. The resources such as the machines and the operators have had an
increase in average utilization. If the results of the deterministic model and the results of
experiment 3 are compared, it can be observed that the machines and the operator
utilization numbers have increased by 10% on average. This is indicative of an optimized
cost structure. The ideal utilization levels for machines and operators may be considered
as 60% - 80%. Excessive utilization of the machine resources may cause excessive wear
and tear and thus higher rate of rejection. Similarly, excessive utilization of the operator
resource may cause fatigue and may introduce defects.
The third constraint was avoiding layoff as much as possible. This constraint can
be attributed to the management’s reluctance to lay off the operators as most of the
57
operators have been working there for a number of years. This constraint was strictly
adhered to. This can be witnessed in the third experiment. In this experiment, the
operators that were busy for a small amount of time were allocated to other tasks. Care
was taken that the workstations where the underutilized operators were originally
allocated do not suffer losses in productivity. Additionally, the new assignments were
designed as to increase the overall productivity of the operation.
This method of productivity improvement is a viable solution for flexibility.
Hypothetically, if the facility needs to make any changes to the facility, either because of
increased demand or because of changes in the product, the facility can implement this
same approach to find the optimized system performance.
Conclusively, we can say that we have fared well with the problem statement that
we were presented with and with the constraints we were supposed to follow. The results
produced through this project, if implemented, can increase the productivity of the facility
by more than 30%. The increase in productivity has little cost associated to it. The cost
burden shared per component is negligible considering the lifespan of the purchased
items, such as conveyors, and the volume of production during that span.
58
Chapter 8
Future Scope
8.1 Scope for Improvement
In this project, the inventory between two stations was reduced from 80 parts to
60 parts. This reduction in inventory levels boosted the production. In future, we will try to
reduce the inventory levels even further. However, there is a possibility that with
increased throughput levels, the material movement system will be overburdened. This
brings us to the second improvement possible. If the material movement system is
overburdened because of the increased throughput, an end-to-end conveyor system
could be installed. This will eliminate the need for manual material movement entirely.
Additionally, we can try to implement a Constant Work In Progress (CONWIP)
system. To successfully implement a CONWIP system, it is necessary that the process is
fairly balanced, with the processing times of each cells being similar. We can observe
from the time study that the process times for the processes vary greatly. Hence, to
successfully facilitate the CONWIP system, we will have to perform extensive line
balancing. One way to achieve this level of line balancing is to introduce automation for
the repetitive processes. Operations from first forming to contouring are repetitive, which
require no special skills from the operator. If these operations are automated, the process
and setup times can be greatly regulated.
The improvements such as end-to-end conveyors and automation require
extensive capital investments. This is not in line with the constraints set forth by the
management for this study. If the constraints were to be relieved, the options such as
conveyors and automation can be explored for an added increase in the productivity.
59
8.2 Scope for Implementation
This project is a turn-key solution for productivity improvement and system
simulation. The resources necessary to commit to and to complete this project are easily
accessible. The data is supposed to be available with the professional working in the
facility. It would require a DES tool that can be purchased online. The technical
knowledge required does not necessarily require years of training. With the
understanding of some of the key concepts of industrial engineering, tasks like the time
study can be successfully performed. The DES tool may need some getting used to but
for an experienced professional, it will not take more than a week. As a result, this project
can be used as a template in any industry similar in size, scope, and operating in similar
market situations.
70
Data fitting and test for normality 1) Shearing setup time
Shapiro-Wilk normality test -
data: x1$Shearing_setup W = 0.96219, p-value = 0.0005338 Pearson chi-square normality test - data: x1$Shearing_setup P = 90.792, p-value = 3.469e-14 Critical Chi Sq. Value = 19.67514 Normal Distribution Parameters - Mean = 0.9802 Standard Deviation = 0.0336 Result = Normality not confirmed.
71
R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. Shapiro Wilk Test - Reference Slawomir Jarek (2012). mvnormtest: Normality test for multivariate variables. R package version 0.1-9. https://CRAN.R-project.org/package=mvnormtest #Normality testing of shearing(Setup_time) x1 <- read.csv(file.choose(),header=T) qqnorm(x1$Shearing_setup,main = "Normal Q-Q Plot: Shearing_setup") qqline(x1$Shearing_setup) shapiro.test(x1$Shearing_setup) z1=pearson.test(x1$Shearing_setup) z1 qchisq(0.05,z1$n.classes-4,lower.tail=FALSE)
79
Value Stream Map
References
[1] Pius Achanga, Esam Shehab, Rajkumar Roy, Geoff Nelder, (2006),"Critical
success factors for lean implementation within SMEs", Journal of Manufacturing
Technology Management, Vol. 17 Iss 4 pp. 460 – 471
[2] Fawaz A. Abdulmalek, Jayant Rajgopal,” Analyzing the benefits of lean
manufacturing and value stream mapping via simulation: A process sector case
study”, Int. J. Production Economics 107 (2007) 223–236
[3] Taho Yang, Yiyo Kuo, Chao-Ton Su, Chia-Lin Hou,” Lean production system
design for fishing net manufacturing using lean principles and simulation
optimization”, Journal of Manufacturing Systems 34 (2015) 66–73
[4] Adam Graunke, Michael Christian, George Thiers, Timothy Sprock, Leon
McGinnis,”AUTOMATED PRODUCTION SYSTEM SIMULATIONS USING
COMMERCIAL OFF-THE-SHELF SIMULATION TOOLS” , Proceedings of the
2016 Winter Simulation Conference.
[5] David J. Meade, Sameer Kumar, Abdolazim Houshyar,” Financial Analysis of a
Theoretical Lean Manufacturing Implementation Using Hybrid Simulation
Modeling”, Journal of Manufacturing Systems Vol 25/No. 2
[6] Pavel Vik, Luis Dias, Ricardo Abreu,” USING SIMIO FOR THE
SPECIFICATION OF AN INTEGRATED AUTOMATED WEIGHING SOLUTION
IN A CEMENT PLANT” Proceedings of the 2010 Winter Simulation Conference
[7] Natalia P. Basán Mariana E. Cóccola Carlos A. Méndez,” OPTIMIZING THE
DESIGN AND OPERATION OF A BEER PACKAGING LINE THROUGH AN
80
ADVANCED SIMIO-BASED DES TOOL” Proceedings of the 2014 Winter
Simulation Conference
[8] J. Ethan Brown, David Sturrock,” Identifying cost reduction and performance
improvement opportunities through simulation”, Proceedings of 2009 winter
simulation conference
[9] C. Dennis Pegden,” Simio: A new simulation system based on intelligent
objects”, Proceedings of 2007 winter simulation conference
[10] Lisa Hawkins,(2001)”Fundamental Productivity Improvement Tools and
Techniques for SME” ISBN 1-84402-001-0
[11] Peter Hines Nick Rich, (1997),"The seven value stream mapping tools",
International Journal of Operations & Production Management, Vol. 17 Iss 1 pp.
46 – 64
[12] Anand Gurumurthy Rambabu Kodali, (2011),"Design of lean manufacturing
systems using value stream mapping with simulation - A case study", Journal of
Manufacturing Technology Management, Vol. 22 Iss 4 pp. 444 – 473
[13] A. Gunasekaran & J. Lyu (1997) Implementation of just-in-time in a small
company: A case study, Production Planning & Control, 8:4, 406-412,
DOI:10.1080/095372897235217
[14] Christer Karlsson Pär Åhlström, (1997),"A lean and global smaller firm?",
International Journal of Operations & Production Management, Vol. 17 Iss 10 pp.
940 – 952
[15] Richard B. Detty & Jon C. Yingling (2000) Quantifying benefits of conversion to
lean manufacturing with discrete event simulation: A case study, International
Journal of Production Research, 38:2, 429-445
81
[16] Surendra M. Gupta & Louis Brennan,” Implementation of just-in-time
methodology in a small company” Production planning and control, Vol.6 1995
[17] http://nptel.ac.in/courses/112107142/part1/table9_1.htm
Biographical Information
Unmesh Vishwas Tamhankar currently lives in Arlington, TX. He is originally from
Karad, India. He has earned his B.E. in Mechanical Engineering from Shivaji University,
India before earning his Master’s Degree (M.S.) in Industrial Engineering from The
University of Texas at Arlington in 2017. Unmesh was the president of the APICS Student
chapter at UTA and was associated with APICS, SWE, NSBE, CSCMP during his time at
the University of Texas at Arlington.